WINNING THROUGH INCREMENTAL INNOVATION: THE CASE OF MYSQL AB

Winning through incremental innovation:
                         The case of MySQL AB

                          Jesper Holck1, Volker Mahnke2, and Roberto Zicari3

Abstract: This empirical paper explores how entrants win substantial market accep-
tance, based on a series of incremental software innovations supported by an open
source software development process, and a unique web-based proprietary business
model. We investigate the competitive processes a case company (MySQL AB) em-
ployed to establish a sustainable foothold in the database software market - against
the odds of a large installed customer base of competitors, and market incumbents’
deep pockets and control of substantial complementary assets. Our findings suggest
that a series of rapid incremental innovations, stack-based complement strategies, an
open-source product testing approach, and a dual licensing strategy all contributed
to winning through incremental innovation in increasingly commoditized database
software markets. We distill seven important lessons that serve to guide managers in
pursuing winning business models based on a series of incremental, and simultane-
ously pursued innovations.
Keywords: Incremental innovation, market share, open source, databases.

1. Introduction

Many current theoretical and practical accounts of entrants winning markets through
innovation are predicated on the crucial assumption that it takes radical, disruptive, or
architectural, rather than incremental innovation to conquer substantial market share
from incumbents, that enjoy network externalities, a substantial installed customer
base, control complementary product markets, and command deep pockets (Utterback
& Abernathy, 1975; Teece, 1986; Abernathy & Clark, 1985; Henderson & Clark,
1990). Indeed, in many industrial contexts, incumbents benefit from incremental inno-
vation. Because of their large installed customer base, incumbents enjoy relatively
greater return from incremental innovation, and also benefit from learning curve ef-
fects. If supply-side economies of scale are present, having more customers usually
translates into greater cost efficiency, that may again be translated into higher margins
and/or lower product prices.
At the same time, incremental innovations are usually thought to enhance an incum-
bent’s capabilities, while radical ones are thought to undermine it. Finally, a large in-
stalled base makes providers of complementary products more likely to provide seam-
less interfaces to a focal incumbent’s product offering. For all these reasons it appears
plausible that winning through incremental innovation against a well-versed set of in-
cumbents is difficult, if possible at all.
However, important exceptions appear to exist. Based on a thorough and longitudinal
case analysis (Eisenhardt, 1989), we identify crucial elements and conditions that
serve to develop an integrative model of winning through incremental innovation in
1
    Copenhagen Business School, Department of Informatics, Howitzvej 60, DK-2000 Frederiksberg, Denmark. jeh.inf@cbs.dk
2
    Copenhagen Business School, Department of Informatics, Howitzvej 60, DK-2000 Frederiksberg, Denmark. vm.inf@cbs.dk
3
  Universität Frankfurt am Main, Institute of Computer Science, Robert-Mayer-Str. 10, D-60325, Frankfurt /M., Germany.
zicari@cs.uni-frankfurt.de

highly competitive markets, against powerful and well established incumbents. While
current innovation theory suggests that winning through incremental innovation
against incumbents is rare and difficult, we suggest that competitive processes em-
ployed by a case company – MySQL AB – serves as a powerful reminder that a
sustainable foothold in the database software market could be established, based on a
series of incremental and complementary innovation.
Despite all odds in the current market, our findings suggest that a combination of such
incremental innovations has enabled competitive moves, including stack-based prod-
uct development strategies, an open source product testing approach, and, important-
ly, dual licensing contracts. These moves all contributed to winning through incre-
mental innovation in the increasingly commoditized database software markets
against powerful competitors such as IBM, Oracle, and Microsoft.
The remainder of the paper is structured as follows. First, we briefly review current
theories on innovative entry, which collectively suggest that winning through incre-
mental innovation is an unlikely industry event. Secondly, we outline our research
methodology and provide a techno-historical account on how MySQL AB entered the
database market and won a sustainable foothold against dominant incumbents such as
IBM, Microsoft, and Oracle. Third, we use the case findings to develop an integrative,
yet parsimonious model of winning through incremental innovation. Fourth, we dis-
cuss alternative explanations of MySQL AB’s success, and, finally, suggest fruitful
avenues of future research.

2. Competitive effects of incremental innovation

Innovations are central in the explanation of competitive success – be they incremen-
tal or radical (Utterback & Abernathy, 1975; Teece, 1986; Abernathy & Clark, 1985;
Henderson & Clark, 1990; Teece, Pisano, & Shuen, 1997). However, many theories
predict that compared to new entrants, established incumbents will benefit most from
incremental innovation4.
Industry life-cycle theories of innovation (e.g. Utterback & Abernathy, 1975; Klepper,
1996) suggest that incumbent firms, engaging in process innovation during and after
the emergence of a dominant design, excel in competition. Thus, in mature industrial
contexts, incremental innovations are likely to benefit incumbent players most. In-
cumbents also enjoy greater returns from incremental innovation the greater their cus-
tomer base (Klepper, 1996): returns are greater if more customers benefit.
In addition, a large installed customer base makes providers of complementary prod-
ucts more likely to offer seamless interfaces to a focal incumbent’s product offering
(Teece, 1982). This is particularly true if such complementarity effects are accompa-
nied by network effects (Liebowitz & Margolis, 1998): while complementary prod-
ucts increase the demand of what an incumbent uniquely provides, network effects in-
crease customer utility as more consumers adopt a focal firm’s product (Dixit & Nale-
buff, 1993)
Such supply-side scale effects work in combination with learning curve effects, fur-
ther augmenting an incumbent’s advantage (Nelson & Winter, 1982). While radical
and/or architectural innovations threaten to undermine an incumbent’s previous
strength (Tushman & Anderson, 1986), incremental innovations usually enhance an
incumbent’s capabilities (Adner & Levinthal, 2001). The scale and learning effects in

4
    For a comprehensive summary see (Tran, 2005)

conjunction provide cost advantages that offer incumbents greater competitive pricing
possibilities, compared to new industry entrants (Henderson & Clark, 1990).
As a result, incumbents in most industries usually win against new entrants if entrants
compete on incremental innovation. It is therefore surprising, that winning through in-
cremental innovation is indeed possible, as illustrated by the following case.

3. The case of MySQL AB

3.1 Methodology

We will in this section present a “techno-historical” account of the company MySQL
and its software products, and use this description as basis for a subsequent analysis
of MySQL’s strategy. Based on this analysis, we will then discuss in what ways the
case of MySQL’s success can be understood in terms of the existing literature on in-
cremental innovation, and finally suggest a few striking “lessons to be learned” from
this case story.

3.2 MySQL: A techno-historical account

MySQL AB was in February 2008 acquired by Sun Microsystems, Inc., for close to
$1 billion in total consideration. This event is, however, just the endpoint of an evolu-
tionary wealth-creating process, made possible by a series of incremental innovation
through which the firm has become a serious competitor and sustained a foothold
alongside IBM, Microsoft, and Oracle, the big three IT companies in the global high-
margin-commodity database business.
MySQL is a SQL database management system (DBMS) with more than 10 million
installations, and approximately 50,000 daily downloads. The basic program runs as a
server, providing multi-user access to a number of databases. Until Sun’s acquisition,
MySQL was under the ownership of a single for-profit firm, the Swedish company
MySQL AB, which also holds the copyright to most of the code base. MySQL AB
was founded by David Axmark, Allan Larsson, and Michael “Monty” Widenius.
Mårten Mickos, the current CEO, joined later as the company grew and sought pro-
fessional management.

1982-1995: Foundation, the first technological battle
MySQL AB’s history goes back to 1982, when MySQL’s co-founder Michael “Mon-
ty” Widenius while employed at TcX, a Swedish IT and data consulting firm, au-
thored UNIREG – a program for administrating ISAM data stores from a text-based
terminal.5 UNIREG was originally written in BASIC for what was at that time state-
of-the-art personal computers with 4 MHz Z80 processors and 16kB of RAM. In the
following years UNIREG was ported to other platforms, from 1986 including Unix.
However, around 1994 UNIREG began to appear outdated, and TcX began to turn its
attention to mSQL (mini SQL) – a proprietary, but inexpensive data base management
system developed in 1994 by Hughes Technologies, which filled an unserved market
need between desktop systems like Microsoft Access and enterprise-level systems like
Oracle and DB2. While mSQL had the advantage of a well-developed client applica-
tion program interface (API), its initial version was not compatible with UNIREG and
also lacked indexing, a feature crucial for performance of large data storage systems.
5
   ISAM (Indexed Sequential Access Method), originally developed by IBM for mainframe computers, is a method for storing
large amounts of data for fast retrieval. ISAM forms the basic data store for almost all databases.

So the idea of simply using mSQL as an alternative front-end for UNIREG was even-
tually abandoned.
Instead, TcX decided to develop a new product, combining the effective UNIREG in-
dexing scheme with an API very similar to mSQL’s. This product was released in
1995 under the name of MySQL version 1. By re-using mSQL’s API, TcX made it
easy to port applications from mSQL to MySQL. mSQL and MySQL were for some
time competitors on the Linux platform, but MySQL eventually won this initial battle,
due to advantages such as fast and efficient code, developed for very small and slow
computers.
The first release of MySQL was not open source, instead TcX chose to release the
software as “shrink wrap”, but with a license that required no payment for deployment
on Unix servers – a license similar to the one used by Aladdin Software for Ghost-
script. The software was freely distributable, and commercial use was allowed as long
as the software was not distributed commercially. Part of MySQL’s initial success is
no doubt based on the choice of a quite “open” license, compared to the one used for
mSQL (Axmark & Widenius, 1999).
It is worth noting that relational databases (RDBMS6) – compared with most other IT
technologies – is a quite old and well-established concept. The basic principles were
developed by Codd as early as 1970 (Codd, 1970); SQL, the standard language for us-
ing relational databases, was first described in 1974, and later adopted as an ANSI
standard in 1986 and as an ISO standard in 1987. So already in the 1980s, the
RDBMS market was technologically quite mature in the sense that customers knew
what to expect of a RDBMS and which competing products were relevant to consider.

1996-2000: Riding on neglected Web service market opportunities
When MySQL.com was formed to maintain the database, its investors requested that
MySQL would be made open source, so they knew the product remained accessible
for use, in case the company would run into trouble. MySQL AB released its product
under the GPL (Gnu Public License).
From 1996 to 2006 MySQL gained an increasingly stable foothold in the database
market by serving the needs of the growing Web service community, initially ignored
by the market leaders (IBM, Microsoft, Oracle). While the initial target was on web-
enabled companies (including Yahoo, Google, Wikipedia, and many smaller ones),
after the foothold was established, MySQL was in the position potentially to chal-
lenge every player in the RDBMS market.
Nonetheless, MySQL had several draw-backs compared to competing products, not
least lack of support for important SQL elements like stored procedures, transactions,
and foreign keys. But in the late 1990s, where “everyone” joined the Internet and
WWW revolution, these draw-backs were of less importance.
The company had already for a long time tried to follow the “15 minute rule”, stating
that it should be possible to get MySQL up and running in less than 15 minutes. This
made MySQL a very attractive choice for Internet start-ups, not wanting to spend
days or weeks on installing complicated RDBMSs and on participating in vendors’
courses necessary to install and administer these systems.
Because most web-administrators had no formal education in computer systems and
database technology, they were less considered with the missing features of MySQL.
They needed the basic RDBMS functionality as a data store for customers, goods,
etc.; unlike traditional enterprise RDBMS users, they were not too worried about e.g.

6
    Relational DataBase Management System

the small risk of an eventual inconsistency in the database, and hence did not feel a
great need for support of ACID transactions.7
The free (both in the sense of gratis and the sense of freedom) availability of the
MySQL software made it an attractive option for start-up companies joining the Inter-
net revolution. Even for established companies, the free availability of MySQL could
be a major advantage. For example, an interesting aspect of a gratis software in large
organizations is that “nerds” in the computer department can test it of, e.g. for an ex-
perimental web site, without asking for financial resources to do this (Holck et al.,
2005). If the software proves to be of good quality, it can then gradually spread in the
organization.
Compared with legacy products like DB2 and Oracle, MySQL had the strong advan-
tage of demanding very little computer resources. Even on inexpensive PCs, MySQL
could run quite efficiently. Partly because of MySQL being open source, a lot of 3rd
party applications (both open source and proprietary) became quickly available, e.g.
for on-line discussion forums and e-commerce. Typically these products were based
on Linux as the operating system; Apache as the web server; MySQL as the database;
and Perl, Python or PHP as the scripting language. As this technology matured, it be-
came known as the LAMP stack8, and attracted lots of seamless 3rd party products and
services, education, and supporting books. Regarding the LAMP stack, a manager ex-
plained to us:
             “MySQL got ignited from the Perl community, through being a better substi-
             tute for mSQL (more stable, more performance). And MySQL grew with the
             PHP community, becoming and staying a default in most PHP scenarios.
             MySQL recently got a default status with Ruby on Rails. Programming lan-
             guage communities are hugely important, when it comes to growing the user
             base”.
Regarding MySQL’s entry on the RDBMS market, he commented:
             “MySQL initially focused on the web-based database market. In a way, that’s
             niche. We conquered an under-served niche with a product aspiring to be
             general-purpose. Then, we’ve conquered related / neighbouring niches. We
             have not grown market share equally in all markets.”
In 1998, Linux Journal announces MySQL as the Readers Choice Award: Favorite
Database, an award MySQL receives every year from 1998 to 2003. The 1999 an-
nouncement said:
             “MySQL, the free database from Finland and Sweden, scores big points with
             42.4%, compared to PostgreSQL with 19.5% and Oracle with 15.2%. [...] it
             stands to reason that MySQL would be a favorite choice of the Linux commu-
             nity, considering its origins and relatively free license.” (Kroll, 2000)

2001-2007: Gaining traction and creating value in increasingly contested
terrain
In 2001, in the aftermath of the Internet bubble, MySQL AB consisted of 20 people
focusing mainly on technology, but largely neglecting managerial issues. In order to
strengthen the company, the founders recruited a number of highly professional peo-
ple for the top management, including chairman of the board John Wattin and CEO
Mårten Mickos. These appointments coincided with the incorporation of the company
in Sweden as MySQL AB. Subsequently, the company showed a strong growth in
7
  ACID = Atomicity, Consistency, Isolation, Durability: a set of properties that guarantee reliable processing of database transac-
tions.
8
    The term was apparently first mentioned under this acronym in an article from 1998 (Kunze, 1998).

employees, operating income, and profits as illustrated below (data from Bureau van
Dijk’s Amadeus database):
                                                            2006                 2002
Operating revenue (thousand SEK)                         285,742                5,611
Total assets (thousand SEK)                              253,934               28,486
Employees                                                    285                    32
Profit (loss) before tax (thousand SEK)                 -115,897              -14,460

This strong growth went hand in hand with establishment of several international of-
fices and subsidiaries, including US, Germany, France, UK, Ireland, Benelux, Japan,
and Italy. MySQL also established partnerships with a large number of companies, in-
cluding NEC America, Sabre, Novell, Cox Communication, SAP, QNX, IBM, Alca-
tel, German Lotto, Sage Group, and Unisys. More professional people were headhunt-
ed, e.g. Kaj Arnö (VP of community relations, 2005). In order to fuel this growth,
venture capital was raised in 2001, 2003 ($19.5 million), and 2006 ($18.5 million).
An important and interesting issue in these years was the “battle of the storage en-
gines”. The MySQL database must be supported by a storage engine, the program that
handles how data are read from and written to the files etc. where the data are stored.
The most common and fastest storage engine for MySQL has probably been
MyISAM. For applications with high demands on concurrency, row-level locking,
and transaction support (ACID compliance), InnoDB has been the preferred choice.
InnoDB is included in the MySQL Enterprise Server and customers wanting to use
MySQL and InnoDB under a commercial license can purchase this through MySQL.
In an interesting move, Oracle Corporation in October 2005 bought Innobase OY, the
company behind InnoDB. Later in February 2006, Oracle increased the pressure on
MySQL, as it also purchased Sleepycat Software, distributor of Berkeley DB, an al-
ternative storage engine.
As both InnoDB and Berkeley DB are open source software, the fact that Oracle Cor-
poration purchased Innobase and Sleepycat did not appear an immediate threat to
MySQL’s business model, but MySQL felt the heat emanating from increasing com-
petitors’ attention. In response, the company engaged with Jim Starkey, creator of
InterBase, a database released by Borland as open source in 2000. In 2007 MySQL re-
leased its first “own” open source storage engine: Falcon.
MySQL was in February 2008 acquired by Sun Microsystems, Inc., for close to $1
billion in total consideration. This event marked an endpoint in the evolutionary
wealth-creating process, made possible by a series of incremental innovation

4. MySQL AB Incremental Innovative Strategy

The development of MySQL does not seem to introduce major technical innovation;
only small incremental improvements to the basic technology were repeatedly imple-
mented. These and the following seven “lessons” for winning through incremental in-
novation can be extracted. Much of MySQL’s effort seems to be have put into making
“behind the scenes” improvements like optimizations, data compression etc.

Lesson #1: Piggyback on existing standards
Always following existing standards and interfaces. By relying on the existing API of
mSQL and the existing ISAM data storage technology, MySQL AB made it easy for

users of these technologies to switch to MySQL. A firm commitment to existing stan-
dards reduces switching costs and helps customers migrating from other suppliers. Do
not let technological advances make your product incompatible with existing soft-
ware.

Lesson #2: Distinguish nice-to-have from must-have-features
Always prioritize a lean product (e.g. fast, easy to install, small memory footprint)
over advanced nice-to-have features. Compared with both commercial and open
source competitors (Oracle, PostgreSQL etc.), MySQL has been quite slow to adopt
features like transactions and foreign keys, otherwise considered “must-have” for pro-
fessional databases. However, as David Axmark and “Monty” Widenius suggest:
“When designing MySQL, we had a greater need for speed than for transactions. It’s
no use having transactions, if the SQL server becomes so slow it’s unusable for what
you need to get done. There are, of course, applications that require transactions to
work, but a wide range of applications work very well without them” (Axmark &
Widenius, 1999)

Lesson #3: Seek supply-side economics of scale by coupling development of
commercial and non-commercial product versions
MySQL chose to let both the Community and the Enterprise version build on the
same code base, i.e. the exact same source code files. In Kaj Arnö’s blog from Octo-
ber 17, 2006, where he introduces the MySQL Enterprise version, he only mentions
one difference between the software offered to the two groups, namely that “we will
do more frequent binary releases of the MySQL Enterprise Server software than of
the MySQL Community Server”, essentially just making it more convenient for En-
terprise users to install the newest version of the software. Also, when a user in the
MySQL forums raised the question of the difference between the Enterprise and the
Community versions, Victoria Reznichenko, support engineer at MySQL answered:
       “They are mostl[y] identical. To the recent Community release was added
       new feature SHOW PROFILE that is not included to Enterprise… InnoDB
       source is the same in both Enterprise and Community, but Enterprise builds
       are made more frequently. So if you hit a bug that [is] already fixed with
       Community binaries you need either to wait till the next Community release or
       build MySQL from source. With Enterprise subscription [y]ou get also access
       to MySQL support and support engineers will help you to find a cause of your
       problem.”
By linking the development of commercial and non-commercial product versions,
economies of scale are realized. This coincides with the high importance that many
open source projects make to maintain one common code base and avoid “forking”; if
the code base is split into parallel streams, it quickly becomes very difficult to transfer
bug-fixes and other code-changes from one stream to another (Holck & Jørgensen,
2003).

Lesson #4: Nurture and coordinate user communities
MySQL makes the following differentiation between the various groups involved in
development and use of MySQL:
      MySQL employees: core developers, often selected from the open source de-
       veloper community;
      developer community: active users, contributing with e.g. code, interfaces, fo-
       rum answers, blog entries;

     user community: including inactive users, silently using the MySQL software.
An important factor for MySQL’s success is the company’s ability to create a single,
coherent community out of the various groups with interests in the product: the free
open source developer community, the community of developers at commercial cus-
tomers, the company’s core developers, etc. All of these developers collaborate using
the Internet in a virtual, coherent community associated with the product and its tech-
nology. The feeling of belonging to this specific community is an important factor
motivating members to take an active role in product development and marketing.

Lesson #5: Use communities to reduce product testing and marketing costs
It is important to note that most members of MySQL’s open source software develop-
er community perform an important support function, but do not generate revenue di-
rectly. With 50,000 product downloads per day and over 10 million active installa-
tions, most community members, even if they are non-paying users, help building a
large community of marketeers and co-developers. While one commercial customer
or even one commercial deal may correspond to hundreds of thousands, or millions,
of installations; open source communities do contribute to reduce product testing and
marketing costs. Community users, e.g., provide very valuable testing, error-report-
ing, and bug-fixing. Indeed, if MySQL chose to make new versions available for
enterprise customers only, these versions would most likely show more errors, be less
stable, and reach acceptable reliability levels later. In addition, because professional
users can toy and play for free with non-commercial versions without asking for
closely guarded IT budgets, they can and often do act as active bridgeheads into
interesting customer accounts (Holck et al., 2005).

Lesson #6: Implement dual-licensing strategy
In contrast to many open source models, MySQL never ceded property rights to de-
veloped software code. So software contributions (e.g. bug-fixes) will only be accept-
ed if the contributor accepts to transfer his/her copyright of the source code contribu-
tion to MySQL. As not all external contributors may want to do this, MySQL runs the
risk of losing contributions that might significantly improve the software’s quality.
But having sole copyright over the source code is a fundamental and necessary pre-
requisite for applying the dual-licensing strategy, where MySQL offers different li-
censing arrangements for different user groups: GPL9 for non-commercial users; com-
mercial license for commercial users. Dual licensing is an important element in bene-
fiting from an incremental innovation based business model for software products that
are embedded in other products. According to Mickos, most of MySQL AB’s revenue
is generated from the 0.5% who pay for embedding MySQL in their proprietary prod-
ucts.
We can identify two categories of paying customers: (a) Enterprise users, paying for
the services included in the MySQL Enterprise subscription; and (b) Re-distributors
buying the software with a non-GPL license and embedding it into their own, com-
mercial product(s). For enterprise users, the dual licensing model is essentially
equivalent to a support model – customers pay for licenses and in return receive sup-
port from MySQL AB. The revenue generated by MySQL Enterprise Subscription10 is
steadily increasing, which can be seen as an indicator that a large community may be
a good base for revenue generation. One factor to consider is a possible large time lag
9
 Gnu Public License, the most popular open source license, implying that all software product derivatives must also be licenses
under GPL (http://www.gnu.org/licenses/, accessed June 2008)
10
  MySQL at present offers four levels of Enterprise Subscription, ranging from Basic (€ 479 per server per year) to Platinum (€
3,999 per server per year). Platinum subscription includes 24x7 support access, 30 minute emergency response time, consultative
support etc. (http://www.mysql.com/products/enterprise/features.html, accessed June 2008)

(of several years) between the first download and becoming a paying customer. How-
ever, a frequent trigger for becoming a commercial MySQL customer is the use of
MySQL for business critical purposes.

Lesson 7: Link the power of market and communities
In addition to providing candidates for viral marketing and product testing, communi-
ties also serve as recruiting ground for commercial customers. Seen this way, winning
through incremental innovation means combining the power of community and mar-
kets, supplying both “Community” and “Enterprise” versions of the database soft-
ware. MySQL exploits the power of the market by offering enterprise customers extra
value for their money (e.g. new features, more functionality, better performance, and
fewer errors), while community users – as previously mentioned – provide very valu-
able testing, error-reporting, and bug-fixing. The most important difference between
the Enterprise and the Community version is not the software but the amount of sup-
port, customers receive from MySQL. As Kaj Arnö describes the key features of the
Enterprise edition:
       The MySQL Enterprise Server is for the non-DIY [do it yourself] commercial
       user; part of the ‘MySQL Enterprise’ subscription offering; for those who
       want extra help developing, deploying and managing MySQL DBs; coupled
       with access to MySQL technical support; assisted by new automated DBA
       monitoring and advisory services. And he continues, The driving force behind
       many commercial relationships involving both normal MySQL Server and
       MySQL Cluster is the mission-criticality of the use. If lots of customer rev-
       enues depend on MySQL working, then the commercial relationship gets go-
       ing.
By combining the power of the market with the power of the community, incremental
innovations pursued simultaneously can become the kernel of sustained competitive
success.

5. Discussion and conclusions

Incumbents in most industries (e.g. IBM, Oracle, Microsoft in the increasingly com-
moditized database market), usually win against new entrants (such as MySQL) if en-
trants compete on incremental innovation. MySQL provides and interesting exception
to this rule: It succeeded not only to establish itself in a highly competitive market but
also directed its value creating evolution into a beneficial acquisition by Sun in the
spring of 2008. This was achieved through a series of incremental and software inno-
vations, accompanied by a series of competitive as well as collaborative actions: (1)
piggybacking on existing standards; (2) pursuing a lean product development strategy,
distinguishing nice-to-have from need-to-have features; (3) seeking supply side
economies of scale by coupling the development processes of commercial and com-
munity versions of the product; (4) actively nurturing communities; (5) using them to
reduce product testing and marketing expenses; (6) applying a dual licensing
strategy ; and finally (7) by combining all these actions into a coherent action portfo-
lio, a business model unleashing the power of both community and markets.
Alternative explanations available in the current literature only partially account for
the ‘winning-through-incremental-innovation’ phenomenon observed in our case. If
an innovation is competence-destroying to industry incumbents, a radical product dis-
continuity or technological shift either creates a new product class (e.g. airlines vs. au-
tomobiles) or substitutes for an existing one (e.g. records versus compact disks).

However, in the case of MySQL neither a radical product discontinuity nor a substan-
tial technological shift was present from a software development perspective.
As an alternative explanation, the phenomenon of disruptive innovation (Christensen,
2003) could be applied. It occurs when neglected market segments are discovered and
served, based on incremental innovation that simplifies product offers in technologi-
cally over-served markets. When such market niches exhibit greater growth rates than
incumbents’ market do, firms populating the niche markets may win in the long run.
While this explanation goes some way to explain the initial success of MySQL in the
web-enabled enterprise market, the software solution offered by MySQL is neither
substantially inferior to possibly over-complicated competing products, nor does the
product offering serve a more rapidly growing main market (e.g. web-based enterpris-
es such as Yahoo, Google etc).
A third, alternative explanation might be that what appears to be an incremental inno-
vation is indeed an architectural innovation (Henderson & Clark, 1990), a change in
the way product components are assembled. If unnoticed by the incumbent, incremen-
tal architectural changes may over time undermine its competitive position. This per-
spective may yield a future research agenda in that a stack-based software product
could be interpreted as architectural design, integrating software functionality across
firm boundaries; this would amount to a possible explanation of one particular aspect
of our case story: MySQL’s orchestrating its offer with other LAMP stack providers.
In sum, while alternative explanations go some way to illuminate partial aspects of
how firms can win through incremental innovation in commoditized markets, such as
the database software market, we believe that the lessons distilled in this paper serve
as a foundation for crafting an alternative theory. The lessons suggest an integrated
action set on how to combine the power of the market with active utilization of com-
munities in an evolutionary, yet incremental innovation path to create shareholder val-
ue.
To conclude, this empirical paper used the case of MySQL to explore how entrants
win substantial market acceptance based on a series of incremental software innova-
tions, supported by an open yet proprietary source software development process, and
a unique web-based business model. Against all odds of a large installed customer
base of competitors, and market incumbents’ deep pockets and control of substantial
complementary assets, our findings suggest that a series of rapid incremental innova-
tions, stack-based complement strategies, an open-source and community-based prod-
uct testing and viral marketing approach, accompanied by a dual licensing strategy, all
contribute to winning through incremental innovation in increasingly commoditized
database software markets. We distill seven important lessons for winning through in-
cremental innovation to guide and stimulate future research on the relation between
innovation and competitive outcomes in contested software markets.

References

Abernathy, W. J. & Clark, K. B. (1985): “Innovation: Mapping the Winds of Creative
   Destruction”, Research Policy, 14: 3-22.
Abernathy, W. J. & Utterback, J. M. (1978): Strategic Management of Technology
   and Innovation, McGraw-Hill: New York.
Adner, R. & Levinthal, D. (2001): “Demand Heterogeneity and Technology Evolu-
   tion”, Management Science, 47(5): 611–628.
Axmark, D. & Widenius, M. (1999): “MySQL Introduction”, Linux Journal, (67).

Christensen, C. M. (2003): The Innovator’s Dilemma: The Revolutionary Book that
    Will Change the Way You Do Business, HarperCollins.
Codd, E. F. (1970): “A Relational Model for Large Shared Data Banks”, Communica-
    tions of the ACM, 13(6).
Dixit, A. K. & Nalebuff, B. (1993): Thinking Strategically: The Competitive Edge in
    Business, Politics, and Everyday Life, W. W. Norton & Company.
Eisenhardt, K. (1989): “Building Theories from Case Study Research”, Academy of
    Management Review, 14 (4): 532-550.
Henderson, R. M. & Clark, K. (1990). “Architectural Innovation: The Reconfigura-
    tion of Existing Product Technologies and the Failure of Established Firms”, Ad-
    ministrative Science Quarterly, 35 (1): 9-30.
Holck, J. & Jørgensen, N. (2003): “Continuous Integration as a Means of Coordina-
    tion: A Case Study of Two Open Source Projects”, Proceedings from ISD 2003,
    Melbourne, Australia.
Holck, J., Pedersen, M. K. & Larsen, M. H. (2005): “Open Source Software Acquisi-
    tion: Beyond the Business Case”, Proceedings from ECIS 2005, Regensburg,
    Gernany.
Holck, J. & Zicari, R. V. (2006): “A Framework Analysis of Business Models for
    Open Source Software Products with Dual Licensing”, CBS/INF Working Paper,
    No. 1, January 2007, Copenhagen Business School, Department of Informatics.
Horstmann, J. (2005): Migration to Open Source Databases, Diploma Thesis, Techni-
    cal University of Berlin.
Klepper, S. (1996) “Entry, Exit, Growth, and Innovation over the Product Life
    Cycle.” American Economic Review, 86 (3): 562-584.
Kroll, J. (2000): “1999 Readers’ Choice Awards”. Linux Journal, (69).
Kunze, M. (1998): “Let There be Light”. C’T, December: 230.
Liebowitz, S. J. & Margolis S. E. (1998): “Network Externalities (Effects)”, New Pal-
    grave Dictionary of Economics and the Law, MacMillan.
Nelson, R.R. & Winter, S.G. (1982): An Evolutionary Theory of Economic Change,
    The Belknap Press of Harvard University Press.
Teece, D. J. (1982): “Towards an Economic Theory of the Multiproduct Firm”, Jour-
    nal of Economic Behavior and Organization 3: 39-63.
Teece, D. J. (1986) “Profiting from Technological Innovation: Implications for Inte-
    gration, Collaboration, Licensing and Public Policy”, Research Policy, 15(6):
    285-305.
Teece, D., Pisano, G. & Shuen, A (1997): “Dynamic Capabilities and Strategic Man-
    agement”, Strategic Management Journal, 18(7): 509-533.
Tran, Y. (2005): “Industrial dynamics and leadership succession: The case of the fash-
    ion industry”. DRUID working paper.
Tushman, M. L. & Anderson, P. (1986): “Technological Discontinuities and Organi-
    zational Environments”, Administrative Science Quarterly, 31: 439-465.
Utterback, J. M. (1994): Mastering the Dynamics of Innovation: How Companies Can
    Seize Opportunities in the Face of Technological Change, Harvard Business
    School Press, Boston, MA.
Utterback, J. M. & Abernathy, W. J. (1975): “A Dynamic Model of Process and Prod-
    uct Innovation”, Omega, 3(6): 639-656.